Changes in Tropical Cyclone Activity Over the Western North Pacific, Downscaled from Ncep/ncar and ERA 40 Reanalysis

In order to derive climate statistics, long and homogeneous time series are needed. Observational data sets ('best track data') of tropical cyclone (TC) activity in the western North Pacific basin show discrepancies in long-term trends derived for the last decades. Therefore an atmospheric regional model (CCLM) is employed to derive alternative data set. For this purpose TC climatology is dynamically downscaled (1948-2011) from two reanalyses: NCEP/NCAR 1 and ERA 40, as discrepancies between them may potentially lead to different results. The reconstructed yearly variability of intense TC numbers yields remarkable agreement with the observed one. However, reconstructed and observed long-term trends (1948-2011) of TC activity differ. While observation-based data sets show rather decadal variability, both simulations demonstrate a strong increase. Further analysis indicated, that upward trends of downscaled TC activity for 1948-2011 is likely influenced by inhomogeneities introduced to reanalyses in 1978 by satellite-based observations. Differences between two CCLM simulations since 1978 point to uncertainties associated with intrinsic features and quality of the reanalyses. Therefore the interpretation of dynamically downscaled reanalyses should be treated with caution when analysing TC activity, especially for the pre-satellite period.